372 



NATURE 



[December ii. i<)19 



to the "Age of Reptiles," which is illustrated by 

 such huge forms as Brontosaurus aiTd Diplodocus, 

 Stegosaurus and Triceratops, which might be a 

 first attempt at a Pachyderm, together with 

 Megalosaurus, Tyrannosaurus, and Iguanodon, 

 besides IMesiosaurus and Ichthyosaurus in the sea, 

 with flying creatures like Fterodactyles and 

 .\rcha:opteryx, half bird, half lizard ; and then 

 dentigerous birds, which pass on to the Kainozoic, 

 and end the present part of the work, which when 

 completed will be a broad survey of the world 

 throughout time. 



Mr. Wells has undertaken a diflicult task, and 

 it is not too much to say tliat no other writer of 

 the present day is so well equipped as he is to 

 bring it to a successful completion. He possesses 

 the rare combination of brilliant literary power 

 with comprehensive and precise knowledge, and 

 this distinctive quality makes his work one in 

 which all intelligent readers will find profit and 

 delight. 



OUR BOOKSHELF. 

 Principles of Electric Spark Ignition in Internal- 

 combustion Engines. By J. D. Morgan. 

 Pp. vii + 88. (London: Crosby Lockwood and 

 Son, 1920.) Price 8.?. 6d. net. 

 In the eighty or so pages of this little book Mr. 

 Morgan gives the result of certain experiments by 

 others and himself to determine the nature of elec- 

 tric spark ignition in internal-combustion engines. 

 A wide circle of readers will feel grateful to the 

 author for providing in this convenient form an 

 account of the more important work along these 

 lines, and for the reference which he provides to 

 the sources from which information in fuller detail 

 may be obtained. Mr. Morgan puts them still 

 further in his debt by the lucidity with which he 

 writes, and his manifest endeavour — almost 

 always successful — to ensure that, even in the 

 more intricate parts of the subject, his phraseology 

 shall be free from the ambiguity which is so often 

 the despair of readers of technical books. 



Mr. Morgan explains most ingeniously and 

 simply his view of the double nature of the spark, 

 its important "capacity " component and the less 

 valuable "inductance" oscillation. l-^xperiment 

 so far has failed to show any effect on the result- 

 ant gaseous explosion, or on the upper and lower 

 limits of richness at which explosion will occur, 

 of change in the size, temperature, energy, or 

 other feature of the spark. Any one individual 

 spark seems to be as good as any other, pro- 

 vided that explosion is caused ; but the apparatus 

 must be unfailing in the succession of sparks 

 which it is designed to provide. 



The book is one which should be in the hands 

 of all who are interested in the scientific side of 

 (he design of internal-combustion engines in their 

 many forms. That further work along these lines 

 will cause spark gaps to be less sensitive than 

 they now are to short circuiting caused by the 

 inevitable gradual loss of insulation is greatly to 

 be hoped. 



NO. 2615, VOL. 104] 



LETTERS TO THE EDITOR. 



[The Editor does not hold h'lnself responsible lor 

 opinions expressed by his correspondents. Neither 

 can he undertake to return, or to correspond with 

 the writers of. rejected manuscripts intended for 

 this or any other part of Nature. No notice is 

 taken of anonymous communications.] 



Gravitation and Light. 



Ji'inrKK ouf<ht just to show the Einstein detlt:ction, 

 for if it pass between two stars a coupU- of diameters 

 of the planet apart, their temporary relative displac - 

 iiuMit will be a " third " of arc, the sixtieth of a second ; 

 and this could be measured with a heliometer. 



Oliver J. Lodc.k. 



Maricmont, Edgbaston, December 6. 



The Deflection of Light during a Solar Eclipse. 



Prok. Anderson suggested in Nailkk of Decem- 

 ber 4 (p. 354) a possible source of systematic error 

 in the determination of the deflection of light at an 

 eclipse, owing to lateral refraction caused by a tem- 

 perature-gradient in the shadow-cone in our atnio- j 

 sphere. Having carefully considered this suggestion, 

 1 feel convinced that the effects of any possible 

 temperature-gradient would be small. 



Taking the height of the atmosphere as 10 mile^, 

 the ray from a star 30' from the sun's centre would 

 traverse a distance of 150 yards in the direction per- 

 pendicular to the shadow-axis whilst passing throuf<h 

 our atmosphere. Prof. Anderson estimates a tem- 

 perature-drop of i/i8 of a degree as required to pro- 

 duce the observed deflection. Thus the lateral tem- 

 perature-gradient must be 1° C. per 1^ miles. The 

 shadow moves over the earth at about 30 miles a 

 minute, so that for a stationary observer the fall of 

 temjjerature would have to be at the rate of 20'^ a 

 minute to produce the observed effect. 



In the case of a single surface of discontinuity, 

 considered for simplicity by Prof. Anderson, the dis- 

 placement by lateral refraction is inversely proportional 

 to the distance from the sun's centre; but this law- 

 does not apply in the actual case of a continuous 

 temperature-gradient. 



It seems possible that the effect might amount to 

 • IS much ;is 1/20 of the Einstein deflection in some 

 cases, and possibly the rather high value found :it 

 Sobral has been increased by this cause. .\t Principe 

 there was no perceptible chanfje of temperature during 

 the eclipse, but the climatic conditions there ire 

 exceptional.' A- S- Eddincton. 



Observatory, Cambridge. 



Proi-. .Vndersgn's letter in Natlrk of December 4 

 raises a point well worthy of 'consideration — that is, 

 the possibilitv of abnormal refraction due to the 

 lowering of temperature in the air by the p.issajie 

 of the shadow cone. I do not, however, think that 

 more than a verv small portion of the efl'ect noted at 

 Sobral could be explained in this way. The shadow 

 ellipse was 194 miles long (direction of motion) and 

 137 miles broad. I have drawn a section in the former 

 direction to scale, taking the height as 20 miles. It 

 is certainlv unnecessary to take it higher, as the 

 temperature of the upper air is unaffected by the 

 passage of the shadow. 



Photographs were taken at Sobral at uniform 

 intervals throughout totality, and all give tolerablv 

 accordant values of the shift. 



